JP2014127734A - Multitrack sound recording apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a multitrack sound recording apparatus, capable of allowing only a musical performance sound to be superimposed even a musical sound is output from a speaker.SOLUTION: A sound signal collected by a microphone MIC is input in a sound recording section 23 via an echo canceller 22. The sound recording section 23 mixes a digital musical sound signal output from a reproduction section 21 and a digital sound signal output from the echo canceller 22, and records the mixed sound. From the sound signal collected by the microphone MIC, the echo canceller 22 removes recurrence component of the musical sound signal output from a speaker SP that recurred through an acoustic space. Since a sound recording/reproducing apparatus 2 removes only the recurrence component from the speaker SP to the microphone MIC as described above, only a musical performance sound or a signing sound is allowed to be superimposed even a musical sound is output from a speaker.

Description

  The present invention relates to a multi-track recording apparatus for recording a new sound as a separate track on an original musical sound while reproducing the musical sound.

  2. Description of the Related Art Conventionally, there has been proposed a multiplex recording apparatus that reproduces musical sound data recorded on a recording medium and mixes and records new sound on the original musical sound data (see, for example, Patent Document 1).

  As shown in Patent Document 1, the multiplex recording apparatus is provided with an audio output terminal for outputting a reproduced musical sound and an audio input terminal for inputting an audio signal by connecting a microphone. . The user performs or sings while listening to the reproduced musical sound as a monitor sound, and performs recording.

  However, if you record while listening to music with a speaker connected to the audio output terminal, the sound output from the speaker is picked up by the microphone, and the original music and new sound are mixed and recorded. There's a problem. Therefore, the user connects headphones to the audio output terminal and performs or sings while listening to the monitor sound through the headphones. Thereby, only a user's performance sound or singing sound is picked up with a microphone, and is inputted into a multiplex recording device. The input performance sound or singing sound is mixed with the original musical sound data and recorded. At this time, the input performance sound or singing sound is recorded as a new track different from the reproduced musical sound. In this way, multitrack recording is performed.

JP 2009-171099 A

  However, wearing headphones for a long time is painful. In particular, the higher the sound quality of the headphones, the larger and heavier the headphones themselves, and the greater the painfulness. Also, since the sound heard with headphones is different from the actual performance sound, it is difficult to perform in accordance with the original musical sound.

  On the other hand, when recording while listening to music with a speaker, there is a problem that music output from the speaker is mixed with a new track.

  Accordingly, an object of the present invention is to provide a multitrack recording apparatus capable of recording only a performance sound or a singing sound as a new track even when a musical sound is output from a speaker.

  The multi-track recording apparatus of the present invention is a multi-track recording apparatus to which a speaker and a microphone are connected, wherein a reproducing means for reproducing a musical sound and an output for outputting a musical sound signal related to the musical sound reproduced by the reproducing means to the speaker. Means, input means for inputting a sound signal picked up by the microphone, and a regression component obtained by regressing the sound related to the musical sound signal output from the output means through an acoustic space from the sound signal input by the input means It is characterized by comprising: a regression sound removing means for removing; and a recording means for recording the sound signal after removing the regression component as a new track superimposed on the musical sound reproduced by the reproducing means.

  As described above, the multitrack recording apparatus can record only the performance sound or the singing sound even when the musical sound is output from the speaker by removing only the regression component from the speaker to the microphone.

  The regression sound removing means estimates the transfer function of the acoustic space, filters the musical sound signal output from the output means with the estimated transfer function, and generates a pseudo regression sound; And a subtracting unit that subtracts the regression sound from the sound signal input by the input unit, and the multitrack recording apparatus stops the adaptive processing of the adaptive filter when the reproduction unit is reproducing the musical sound. It is preferable to further comprise a control means.

  In this configuration, the regression component is suppressed by the adaptive filter that estimates the transfer function of the acoustic space from the speaker to the microphone. In the transfer function estimation method, when a signal after subtraction is used as a reference signal, a filter coefficient is calculated so that the reference signal becomes as small as possible based on a musical sound signal output to a speaker. Here, the reference signal includes only the regression component that could not be suppressed, which is an ideal state for accurately estimating the transfer function. However, since the user of the multitrack recording apparatus performs or sings while listening to the musical sound output from the speaker as the monitor sound, the sound signal input from the microphone includes the signal of the performance sound or singing sound. In this case, since the reference signal input to the adaptive filter includes signals other than the regression component, if the adaptation process is continued, the filter is adapted as a filter different from the transfer function of the actual acoustic space, and the regression is sufficiently performed. The component may not be able to be suppressed. Therefore, the multitrack recording apparatus stops the adaptive processing of the adaptive filter on the assumption that the performance sound or the singing sound is also input when the reproduction sound is output.

  In particular, it is desirable to output a musical sound (for example, white noise) from the speaker and identify the filter coefficient of the adaptive filter in the transfer function of the acoustic space before stopping the adaptive processing. This makes it possible to perform multitrack recording while maintaining an ideal filter coefficient and suppressing only the regression component.

  According to the present invention, it is possible to superimpose only a performance sound or a singing sound as a new track even when a musical sound is output from a speaker.

It is the schematic which shows the structure of a multitrack recording system. It is a block diagram which shows the structure of a recording / reproducing apparatus. It is a block diagram which shows the structure of an echo canceller. It is the flowchart which showed operation | movement of the control part. It is a block diagram which shows the structure of the recording / reproducing apparatus provided with the band-limiting filter.

  FIG. 1 is a schematic diagram showing the configuration of the multitrack recording apparatus of the present invention. The multitrack recording system shown in FIG. 1 includes a recording / reproducing apparatus 2, a speaker SP connected to the recording / reproducing apparatus 2, and a microphone MIC.

  The recording / reproducing apparatus 2 corresponds to the multitrack recording apparatus of the present invention. The recording / reproducing apparatus 2 reproduces a musical sound and outputs a musical sound signal related to the reproduced musical sound to the speaker SP. The user (performer) performs the musical instrument 1 while listening to the musical sound output from the speaker SP as a monitor sound. Alternatively, the user sings instead of playing the instrument 1 (or playing the instrument 1). The microphone MIC collects ambient sounds and outputs a sound signal related to the collected sounds to the recording / reproducing apparatus 2.

  FIG. 2 is a block diagram showing the configuration of the recording / playback apparatus 2. The recording / playback apparatus 2 includes a playback unit 21, an echo canceller 22, a recording unit 23, a control unit 24, an operation unit 25, a D / A conversion unit (D / A in the drawing) 26, and an amplification unit (AMP in the drawing) 27. , An amplification unit (AMP in the figure) 28 and an A / D conversion unit (A / D in the figure).

  When the user uses the operation unit 25 to instruct the reproduction of musical sounds, the control unit 24 instructs the reproduction unit 21 to reproduce musical sounds and instructs the recording unit 23 to record.

  The reproduction unit 21 reproduces a musical sound that the user listens to as a monitor sound in accordance with an instruction from the control unit 24. For example, the playback unit 21 reads audio data (encoded data such as PCM data and MP3) stored in a built-in memory (not shown), and generates a digital musical sound signal. The digital musical tone signal is converted into an analog musical tone signal by the D / A converter 26 via the echo canceller 22. The analog musical sound signal is amplified by the amplification unit 27 and output to the speaker SP.

  The musical sound output from the speaker SP is collected by the microphone MIC through the acoustic space. Also, the performance sound of the user playing the musical instrument 1 is collected by the microphone MIC. The analog sound signal related to the sound collected by the microphone MIC is amplified by the front-end amplifier 28 and converted to a digital sound signal by the A / D converter 29. The digital sound signal is input to the recording unit 23 via the echo canceller 22. The recording unit 23 mixes and records the digital musical sound signal output from the reproducing unit 21 and the digital sound signal output from the echo canceller 22. At this time, the digital sound signal input from the echo canceller 22 is recorded as a new track different from the reproduced digital musical sound signal. In this way, multitrack recording is performed.

  FIG. 3 is a block diagram showing the configuration of the echo canceller 22. The echo canceller 22 includes a filter unit 221, a coefficient update unit 222, and a post processor 223.

  The filter unit 221 is a digital filter (FIR filter) having a predetermined filter coefficient, filters the digital musical sound signal input from the reproduction unit 21, and simulates a pseudo regression sound that simulates a regression component from the speaker SP to the microphone MIC. Is generated. The filter coefficient of the filter unit 221 is an estimation of the transfer function of the acoustic space from the speaker SP to the microphone MIC.

  The filter unit 221 only needs to have the number of taps that can cover a frequency band (for example, a band of 50 Hz to 20 kHz) that can be output by the speaker SP. Ideally, the filter unit 221 is a digital musical sound signal output from the reproduction unit 21. It is preferable to have a corresponding number of taps. For example, when the sampling frequency of the digital musical sound signal output from the playback unit 21 is 48 kHz, the filter unit 221 covers the frequency band of several tens of Hz to 24 kHz among the direct sound from the speaker SP to the microphone MIC. More than 1000 taps are required. In order to simulate indoor reflected sounds and reverberant sounds, more taps are required.

  Note that the musical sound output from the speaker SP is for the user to listen to as monitor sound, and does not need to have high sound quality. For example, it is sufficient to cover a frequency band of about 200 Hz to 10 kHz. Therefore, like the recording / reproducing apparatus 1A of FIG. 5, a band limiting filter 51 is connected in front of the echo canceller 22, the digital musical sound signal input from the reproducing unit 21 is limited to a predetermined band, and the number of taps of the filter unit 221 is reduced. It is also possible to reduce.

  The pseudo-regression sound generated by the filtering process by the filter unit 221 is output to the post processor 223. The post processor 223 subtracts the pseudo regression sound from the sound (the digital sound signal) collected by the microphone MIC.

  In this way, the regression component from the speaker SP to the microphone MIC among the sounds collected by the microphone MIC is suppressed. The digital sound signal after the regression component is suppressed is output to the recording unit 23 and also to the coefficient update unit 222.

The coefficient updating unit 222 receives the digital musical sound signal output from the reproducing unit 21 and the digital sound signal output from the post processor 223, estimates the transfer function of the acoustic space, and updates the filter coefficient of the filter unit 221. . That is, the coefficient updating unit 222 estimates a filter coefficient that makes the reference signal as small as possible using the digital signal after the regression component is suppressed as the reference signal. As an estimation algorithm, for example, an LMS method can be applied. In the LMS method, when the reference signal is e (k), the mean square value J = E [e (k) ² ] (where E [•] is an expected value) of the e (k) is minimized. A simple filter coefficient is obtained by calculation. In this manner, the pseudo regression sound output from the filter unit 221 can be approximated to the regression component.

  Here, if only the regression component that could not be suppressed is included in the reference signal, after a certain amount of time has passed, the filter coefficient of the filter unit 221 is identified as the transfer function of the actual acoustic space, and the pseudo regression sound is actually It almost agrees with the regression component of. In this case, only the regression component from the speaker SP to the microphone MIC can be removed from the sound collected by the microphone MIC. Therefore, the recording / reproducing apparatus 2 can record only the performance sound of the musical instrument 1 even when the musical sound is output from the speaker SP. Therefore, only the performance sound of the musical instrument 1 can be recorded on the new track. Thereby, the user can perform multitrack recording while listening to the monitor sound without wearing headphones.

  However, the reference signal includes not only the regression component that cannot be removed but also the performance sound of the musical instrument 1. In this case, since the reference signal includes a signal other than the regression component, if the above adaptive processing is continued, a filter coefficient different from the transfer function of the actual acoustic space is set in the filter unit 221. The regression component may not be sufficiently suppressed.

  Therefore, the recording / playback apparatus 2 inputs the performance sound to the microphone MIC when the musical sound reproduced from the speaker SP is output in order to maintain the ideal filter coefficient after a certain amount of time has passed. As an example, the adaptive processing of the adaptive filter is stopped.

  FIG. 4 is a flowchart showing the operation of the control unit 24. First, the control unit 24 determines whether or not the user has operated the operation unit 25 to give a recording / playback instruction (s11). This determination is repeated until a recording / playback instruction is given.

  When receiving the recording / playback instruction, the control unit 24 performs initial measurement (s12). The initial measurement is a process for causing the reproduction unit 21 to output a measurement sound such as white noise and bringing the filter coefficient of the filter unit 221 closer to an ideal state.

  Thereafter, the control unit 24 causes the coefficient updating unit 222 to stop updating the filter coefficient, fixes the filter coefficient of the filter unit 221 and starts playing the musical sound (s13). Then, when the control unit 24 is instructed to stop the reproduction from the operation unit 25 or the reproduction of the musical tone is finished, it is determined that the user's performance is finished (s14). The updating of the filter coefficient of the coefficient updating unit 222 is resumed, and the filter coefficient of the filter unit 221 is released (s15).

  Note that the measurement sound used in s12 is not limited to white noise, and for example, a musical sound instructed to be played back by the user may be used. For example, the control unit 24 causes the reproduction unit 21 to output a musical sound instructed to be reproduced by the user for a predetermined time. Thereafter, in s13, the control unit 24 fixes the filter coefficient and outputs the musical sound instructed to be reproduced from the beginning. Alternatively, the control unit 24 may reproduce the musical tone immediately when reproduction is instructed by the user, and stop the adaptation process after a predetermined time has elapsed.

MIC ... Microphone SP ... Speaker 1 ... Musical instrument 2 ... Recording / reproducing apparatus 21 ... Reproducing unit 22 ... Echo canceller 23 ... Recording unit 24 ... Control unit 25 ... Operation unit 26 ... D / A converter 27 ... Amplifying unit 28 ... Amplifying unit 29 ... A / D converter

Claims (3)

A multitrack recording device to which a speaker and a microphone are connected,
Playback means for playing music,
Output means for outputting to the speaker a musical sound signal related to the musical sound reproduced by the reproducing means;
Input means for inputting a sound signal picked up by the microphone;
A regression sound removing means for removing a regression component obtained by regressing the sound related to the musical sound signal output from the output means through an acoustic space from the sound signal input by the input means;
Recording means for recording the sound signal after removing the regression component as a new track superimposed on the musical sound reproduced by the reproducing means;
Multitrack recording device with The multi-track recording apparatus according to claim 1,
The regression sound removing means estimates the transfer function of the acoustic space, and performs an adaptive filter that generates a pseudo-regression sound by filtering with the estimated transfer function of the musical sound signal output from the output means;
A subtractor for subtracting the pseudo-regressive sound from the sound signal input by the input means;
With
A multi-track recording apparatus, further comprising control means for stopping adaptive processing of the adaptive filter when the reproduction means is reproducing the musical sound. The multi-track recording apparatus according to claim 2,
The multi-track recording apparatus according to claim 1, wherein the controller is configured to output a musical sound from the speaker and identify a filter coefficient of the adaptive filter as a transfer function of the acoustic space before stopping the adaptive processing.

Images